Artem Mats

@Reederey

Coding with Local AI and LLMs: 2 x DGX Spark cluster experiments.

USA

github.com/Reederey87

Replies Reposts Media Articles

Pinned
Artem Mats
@Reederey
Sep 13
DeepSeek-V4.1-Flash REAP-256E running on 2× DGX Spark (GB10): 25.65 tok/s decode, 841,973-token KV pool. Fully reproducible. This is FP4, not a 3-bit or 2-bit EXL3 re-quant. What "REAP-256E" means is the routed-expert pool was pruned 384 → 256 per layer using Cerebras' REAP
GitHub - Reederey87/deepseek-v41-flash-reap-2x-dgx-spark: Serve DeepSeek-V4.1-Flash REAP-256E...
From github.com
3
Artem Mats
@Reederey
17h
DeepSeek V4.1 Flash leads on agentic coding and core harness benchmarks, GLM-5.3-Flash edges out general/enterprise Elo scores, and Xiaomi's MiMo-V2.6-Flash dominates security/cyber tasks.
Xiaomi MiMo
@XiaomiMiMo
19h
Introducing Xiaomi MiMo-V2.6 — Pro & Flash. Frontier intelligence, all the modalities, built in public. 🔹 Two omnimodal models, advancing through scaled reinforcement learning 🔹 Pro performs on par with Claude Opus 5 and GPT-5.6 Sol across most agent benchmarks 🔹 Pro scores
Artem Mats
@Reederey
21h
Grok 4.7 has been released. It offers a noticeable boost while maintaining the same price and speed. Musk had previously stated that Grok 4.7 would be a new pre-trained model with 2.1 trillion parameters, but plans evidently changed. Meanwhile, training for Grok 4.8—with its 2.5
Artem Mats
@Reederey
Sep 12
These guys are doing an awesome project!!! I have it on my machine.
DHH
@dhh
Sep 11
We've officially incorporated Omarchy M: The team bringing Omarchy to Apple M chips and vintage machines alike. Awesome progress already! omarchy.org/news/2026/09/i…
Artem Mats
@Reederey
Sep 12
LibertAIDAI/DeepSeek-V4.1-Flash-REAP-256E Ready-to-go recipe to deploy the DeepSeek v4.1 Flash on your 2xDGX Spark cluster. Thanks @Libertai_DAI - amazing job. huggingface.co/LibertAIDAI/De… #DGXSpark #LocalAI #vLLM #GB10
LibertAIDAI/DeepSeek-V4.1-Flash-REAP-256E · Hugging Face
From huggingface.co
3